Data Banks have been developed in Stroke and Traumatic Coma. Organizing the principles for establishing a data bank which describes a neurologic condition requires proposing and testing entirely new concepts of data management. This study analyzes the underlying organizational and methodological principles which are necessary for optimal functioning of a data bank. The methodology employed by the data banks includes innovative techniques, such as interactive, on-site data entry, and local edit checking of data. Since the data banks consist of multiple clinical centers, which collaborate and pool data, stringent techniques are required to ensure consistent data collection. A data base management system is necessary to handle the hundreds of question parameters involved. Analysis of organizational principles is continuing. Information on this methodology is being disseminated by presentations at seminars, meetings and conferences. A Stroke Data Bank Workshop was sponsored at the Stroke Council of the American Heart Association in 1984.